Regex to parse a list of items

Question

I have a string like (Name := Sam&&Age=:17&& Score:=C6) and a class called Person with around 30 properties - Name, Age, Score, etc. How do I write a regex to parse the string, provided there is a variable number of properties?

Also I was thinking of doing some string replacements to convert the string to a proper JSON and then parse it. Does it sound like a good idea for production code?

I'm using C# 2010.

Is it a typo that you have := in on place and then =: in another ? — Babak Naffas, Nov 12 '11 at 00:39
No, someone has manually edited the DB with inconsistent spaces and equals. — Lincoln, Nov 12 '11 at 00:43

score 1 · Answer 1 · edited May 23 '17 at 10:24

1

You should definitely go with the JSON approach. It is much cleaner and robust than relying on a regex.

Take a look at this: Parse JSON in C#

edited May 23 '17 at 10:24

Community

1
1

answered Nov 12 '11 at 00:37

CodeZombie

5,367
3
30
37

There are a few, like System.Web.Script.Serialization.JavaScriptSerializer. – Lincoln Nov 12 '11 at 00:40
Thx for patching my .NET knowledge, I really have to take a look at these. Thx! – CodeZombie Nov 12 '11 at 00:42

score 1 · Answer 2 · answered Nov 12 '11 at 00:38

1

If your separator is &&, then I'd do String.Split( "&&", myTextString) when use the regex (\w+)\s*:=\s*(\w+) to isolate the key/value pairs.

answered Nov 12 '11 at 00:38

Babak Naffas

12,395
3
34
49

Jacob · Accepted Answer · 2011-11-12T15:59:34.750

This regex should match your input string.

\(\s*((?<PropertyName>.*?)\s*((:=)|(=:))\s*(?<PropertyValue>.*?)\s*(&&)?\s*)*\)

Here's what it means:

\(                          Open paren
\s*                         Optional whitespace
(
    (?<PropertyName>.*?)    Property name group
    \s*                     Optional whitespace
    ((:=)|(=:))             Literal ':=' or '=:'
    \s*                     Optional whitespace
    (?<PropertyValue>.*?)   Property value group
    \s*                     Optional whitespace
    (&&)?                   Optional '&&' (the last one won't have one)
    \s*                     Optional whitespace
)*                          The proceeding can repeat 0-many times
\)                          Close paren

With this, you can do a match against your string in C#:

var regex = new Regex(
    @"\(\s*((?<PropertyName>.*?)\s*((:=)|(=:))\s*(?<PropertyValue>.*?)\s*(&&)?\s*)*\)");
var match = regex.Match(yourString);

Then loop through each property/value pair, setting the properties on your object. Setting the object properties will require some reflection and different code based on the object property types:

var properyNames = match.Groups["PropertyName"].Captures;
var properyValues = match.Groups["PropertyValue"].Captures;
var numPairs = propertyNames.Count;

var objType = yourObj.GetType();
for (var i = 0; i < numPairs; i++)
{
    var propertyName = propertyNames[i].Value;
    var theValue = propertyValues[i].Value;

    var property = objType.GetProperty(propertyName);
    object convertedValue = theValue;
    if (property.PropertyType == typeof(int))
        convertedValue = int.Parse(theValue);
    if (property.PropertyType == typeof(DateTime))
        // ....
    // etc....

    property.SetValue(yourObj, convertedValue, null);
}

Thanks Jacob, very nice :) What do you think about the conversion to valid JSON, looks like the code will be shorter, but I'm not sure about performance and reliability? — Lincoln, Nov 12 '11 at 01:01
I wouldn't bother with converting to JSON before parsing; would be some extra unnecessary work, and it can be tricky (you have to deal with escaping some characters). — Jacob, Nov 12 '11 at 16:03

Regex to parse a list of items

3 Answers3